Repository navigation
Conversation
QualityMax ReviewVerdict: ACTION REQUIRED · Confidence: degraded — required review evidence is incomplete Re-review: Open the QualityMax check and select Re-review after it completes. Files eligible: 18 · Files reviewed: 0 · Files with findings: 0 · Findings: 0 · Inline cards: 0 Priority findings
Review gates
Important files
Review lifecycleUse the inline cards to inspect evidence and suggested remediation. Re-run the QualityMax review after pushing a fix; unchanged cards are identified by their stable finding marker. Dismiss with a reason through the existing QualityMax/GitHub review feedback flow. 0 prior card(s) are stale/resolved on this head. Proof legend: VERIFIED independently judged patch · REPRODUCED verified finding · GROUNDED deterministic evidence · MODEL-ONLY model judgment. QualityMax project results are available in the configured project. Receipt · commit |
![qualitymaxapp[bot]](https://avatars.githubusercontent.com/in/2748338?s=60&v=4){kind=small}
|
| Gate | Result |
|---|---|
| 🔍 AI diff review | |
| 🔍 SAST | completed · 30 eligible / 30 reviewed · claude-haiku-4-5-20251001 |
| 🔍 Canonical PR review delivery | completed · 0 eligible / 0 reviewed · exact-head review #5491666872 and overview #6002650534 confirmed |
| 🧪 Repo Tests | ✅ 889/889 passed (go) |
| 🤖 AI Tests |
Powered by QualityMax — AI-Powered Test Automation
- Record the checkpoint digest each native CLI session received; resend only when the session is new, the checkpoint changed, or entries are being transferred, so it no longer accumulates in CLI history per turn. - Print a /handoff reject reminder after a backend switch that follows unrecorded work, noting interrupted turns and failed tool calls where the backend exposes them (CC is_error, Codex exit codes, OpenCode/agy errors, built-in JSON errors). It never records a rejection itself. - Accept any whitespace after /handoff and its subcommand; reject look-alikes like /handoffs instead of sending them to the model. - Support \| for a literal pipe in the approach and echo the parsed approach, warning when the evidence contains a pipe.
Dropping the oldest entries re-encoded the whole history once per dropped entry. Search for the fewest drops that fit instead: ~800 untransferred entries took ~13s before the backend started and now take under 2s. The test checks the result matches the previous linear search.
eval/handoff runs the live experiment from docs/HANDOFF.md: a fixed Go fixture where raising ReconnectTimeout is the tempting but wrong fix, a scripted phase A with real captured test output, checkpoint vs. control arms that differ only by the /handoff note, optional padding that pushes phase A out of the inline handoff, and one backend-B turn through RunCLI. Scoring samples the timeout constant during the turn (catching reverted attempts), runs hidden tests with the original test restored, and writes per-run artifacts plus summary.md. 'run' prints the plan and only calls backends with --yes; 'selfcheck' verifies the fixture without backends.
1 participant
Long or compacted conversations can bury a rejected approach even when the shared transcript retains its evidence. Add
/handoff reject <approach> | <evidence>to keep an explicit user decision visible to the next backend. For example, a note that increasing the reconnect timeout failed at both 5s and 60s accompanies the next turn after switching providers.The checkpoint is included in every CLI handoff, including native resumes with no missing transcript entries, and in built-in provider system prompts. It survives compaction and session save/resume.
/handofflists entries;/handoff forget <number>removes a rejection, and/clearresets the checkpoint. Removed notes remain superseded by the current list.Security: notes use existing retained-transcript redaction, are capped at 12 KiB, and are identified as user-supplied data. Input recall keeps only
/handoff, avoiding retention of raw pasted evidence. Notes make no additional model request and do not inspect credentials or provider-private state.This first version requires explicit user decisions. Failures are not automatically classified, and model compliance is not enforced.
docs/HANDOFF.mddescribes a controlled live backend comparison; that evaluation has not yet been run.Validation:
go test ./...go vet ./...go run github.com/golangci/golangci-lint/cmd/golangci-lint@v1.64.8 run-ldflags='-s -w -X main.Version=dev'(artifact under.context).